Photographic Image Quality:
Size (physical dimensions, with names for some sizes)
Bigger size is usually “better,” for reasons that we’ll discuss when we get to pixels below. Here
are four of the most popular, although there are other, in-between, sizes that are mostly unique to
Full frame (FF), based on the traditional 35mm film camera, approximately 24×36mm, or 864mm².
This format is found only in a few high-end professional cameras such as the Canon 5DmkII, but it also
provides the point of reference for describing lens length in mm. (For example, a “normal” FF lens
is 50mm; to find the same field of view on a smaller sensor, a “crop factor” or FOVCF is applied.)
APS-C, named after a long-defunct camera type that used standard 35mm film but
shot images that were approximately half the standard size. There are several variations on this, but the most common size is
23.67×15.7mm, or 370mm². These have an FOVCF of 1.5, meaning that a 35mm lens mounted on an
APS-C body becomes the equivalent of a FF “normal” lens at 52.5mm. The vast majority of popular DSLR
cameras use this format; it is now beginning to be used in mirrorless system cameras such as the Sony NEX-7,
which provide a much smaller body while keeping the large-sensor image quality.
Micro four-thirds, 17.3×13mm, or 225mm². This format has been
adopted by Olympus and Panasonic for their mirrorless system cameras. It is a compromise that permits
cameras and lenses to be made smaller while still maintaining very good image quality. FOVCF is 2, thus
a 25mm lens is “normal” in this format.
Compact camera sensors vary in size, but a popular one is 1/2.3 (the diagonal
size in inches) or about 28mm², vastly smaller than the others with an FOVCF of 5.6!
Confusingly, some cameras that are physically larger (sometimes known as “mega-zooms”) might
still use a sensor this small. Buyer beware.
Pixel count (expressed in mega-pixels)
Mega-pixel (Mp) numbers, by themselves, are meaningless—one of the more misunderstood
and abused terms in the camera business. Here’s why.
▸ Each sensor pixel is a discrete device that collects light energy, converts it to electrical
energy, and reports how much it has collected to the camera’s computer chip—which turns it into a digital
value between 0 and 255 (higher means more light or brighter). The final image file is simply a collection
of all the pixel values, with some in-camera processing applied. (Actually, it’s much more complicated
than this because the image is in color, but those details aren’t needed here.)
▸ The range of light values that human eyes can distinguish, from dark to light, is far greater
than any sensor pixel can record. Any shadow value darker than the sensor can “see” is simply
recorded as solid black, and any highlight value brighter than the sensor can accept is recorded as blank white.
The physical size of the pixel determines how much light it can accept. Assuming the same underlying
sensor technology, a larger pixel will always be able to record more gradations between light and dark
than a smaller pixel can. With fewer gradations, the smaller pixel is more likely to report bogus values,
seen as “noise” especially in darker areas of the final image.
▸ So more pixels are not always “better.” Given the same physical size
sensor, more pixels do provide a sharper appearance to the image, but fewer pixels provide smoother light
response. Given two different size sensors with the same pixel count, the larger one will always
provide better response to light. The intended use of the image is also important. A 21-Mp full-frame
file will produce astounding prints at 20"×30" or even 30"×45", but an image on
the Internet can be produced from fewer Mp than any smart phone's built-in camera.